Upgrading from llvm 2.8 to 3.0/3.1

Hi,

I am working on upgrading our compiler backend from llvm 2.8 to 3.0/3.1. I have been reading the release docs, llvm blogs etc that talk about the type system rewrite, exception handling rewrite, new register allocators etc to know the changes. One of our major concerns is that even though our compiler’s run time has boosted a lot, thanks to llvm, our compile time is pretty high. I searched for info related to compile time and run time improvements for the releases but couldn’t find one. It would be great if one could point me out to these docs, if any.

-Chethan

I am working on upgrading our compiler backend from llvm 2.8 to 3.0/3.1. I have
been reading the release docs, llvm blogs etc that talk about the type system
rewrite, exception handling rewrite, new register allocators etc to know the
changes. One of our major concerns is that even though our compiler's run time
has boosted a lot, thanks to llvm, our compile time is pretty high. I searched
for info related to compile time and run time improvements for the releases but
couldn't find one. It would be great if one could point me out to these docs,
if any.

  I don't know we have such information on the release note. You can post your
result, and let others comment on the higher compile time however.

Regards,
chenwj